OpenCL 中全局内存的并发更新 (x += a) [英] Concurrent updates (x += a) to global memory in OpenCL

问题描述

我正在 OpenCL 内核中执行以下操作(简化示例):

I'm doing the following in an OpenCL kernel (simplified example):

__kernel void step(const uint count, __global int *map, __global float *sum)
{
    const uint i = get_global_id(0);
    if(i < count) {
        sum[map[i]] += 12.34;
    }
}

这里，sum 是我想计算的一些数量(之前在另一个内核中设置为零)，map 是整数 i 的映射> 到整数 j，这样多个 i 可以映射到同一个 j.

Here, sum is some quantity I want to calculate (previously set to zero in another kernel) and map is a mapping from integers i to integers j, such that multiple i's can map to the same j.

(map 可以在常量内存中而不是全局内存中，但似乎我的 GPU 上的常量内存量非常有限)

(map could be in constant memory rather than global, but it seems the amount of constant memory on my GPU is incredibly limited)

这行得通吗?一个+="是原子方式实现的，还是并发操作有可能相互覆盖?

Will this work? Is a "+=" implemented in an atomic way, or is there a chance of concurrent operations overwriting each other?

推荐答案

这行得通吗?一个+="是原子方式实现的，还是并发操作有可能相互覆盖?

Will this work? Is a "+=" implemented in an atomic way, or is there a chance of concurrent operations overwriting each other?

这是行不通的.当线程访问其他线程写入的内存时，您需要显式地求助于原子操作.在这种情况下，atomic_add.

It will not work. When threads access memory written to by other threads, you need to explicitly resort to atomic operations. In this case, atomic_add.

类似于:

__kernel void step(const uint count, __global int *map, __global double *sum)
{
    const uint i = get_global_id(0);
    if(i < count) {
        atomic_add(&sum[map[i]], 1234);
    }
}

这篇关于OpenCL 中全局内存的并发更新 (x += a)的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！